Carbohydrate-protein interactions are essential for a wide range of biological processes such as inflammation, bacterial and viral adhesion, and metastasis. As a result, there has been significant interest in identifying carbohydrate binding proteins and developing ligands to modulate their activity. Analysis of carbohydrate-protein interactions is complicated by a number of factors. First, carbohydrates are extremely difficult to isolate or synthesize. Therefore, only small amounts can be obtained in many cases. Second, traditional methods for studying binding are not high-throughput and require large amounts of material. Third, monovalent interactions between carbohydrates and proteins are notoriously weak. To achieve a high avidity, carbohydrate-binding proteins contain multiple binding sites and form multivalent complexes. As a result, carbohydrates must be presented in a multivalent fashion and the spacing and orientation of the carbohydrates play a critical role in recognition. Carbohydrate microarrays contain many different carbohydrates, or glycans, immobilized on a solid support in a spatially-defined arrangement. Using an array, one can evaluate binding of a lectin, antibody, cell, or virus to all the carbohydrates on the array simultaneously. The array provides a multivalent presentation of carbohydrates, and the miniaturized format requires only picogram amounts of each carbohydrate for an assay. We have successfully generated a carbohydrate microarray and assay to detect binding. To construct our array, we attach carbohydrates to a carrier protein, such as bovine serum albumin, to produce neoglycoproteins. These neoglycoproteins, as well as natural glycoproteins, are printed onto epoxide coated microscope slides using a robotic microarrayer to produce microarrays. Neoglycoproteins can be used as multivalent reagents for other assays such as ELISAs, as multivalent inhibitors, and as multivalent immunogens. Recently, we have focused on expanding and improving the carbohydrate microarray. One objective is to increase the number of glycans on the array. Over the last year, we have expanded the size of our array from 125 to 204 array components. Since spacing and orientation of carbohydrates on the array surface is critical for formation of a multivalent complex, a second objective is to explore methods for varying presentation of our array components. We have varied the carbohydrate density by modulating the number of carbohydrates attached to each molecule of albumin. We have shown that this method for varying density is useful for evaluating recognition of lectins, monoclonal antibodies, and serum antibodies. We are currently investigating varying the density of neoglycoproteins by mixing or doping in unmodified albumin.